Re: USAGE: Circumfixes

--- In conlang@yahoogroups.com, Tamas Racsko <tracsko@F...> wrote:
> On 11 May 2004 Richard Wordingham <richard.wordingham@N...> wrote:
>
> > In which case we're back to Ce- -[h/k]-a- as a single morpheme!
(At
> > least, as single as French ne..pas, ne..jamais etc. are.
>
>   I'm sure that in agglutinative languages a circumfix may express
> only one thing: person, tense, mood, voice, aspect, word derivation
> etc.
>
>   In this context I don't know what can be a circumfix in a
> flexional language. I think that we must be careful when a
> flexional affix encodes more grammatical cathegory, e.g. in case of
> ne..jamais: negating the verb (ne = not) and a negative time
> complement (jamais = never). Is English colloquial "ain't
> (do/say/etc) nothing" a circumfix? Is English phrase "and ...
> respectively" is a circumfix?
>
>   I think this is the same for Greek "Ce- -[h/k]-a-". What is the
> difference between the conjunction of allomorphemes and the
> circumfix in flexional languages?
>
>   Are Slovak active past participles -- like u.vide.vs^i' 'seen,
> seeing' < vidie.t' 'to see' -- formed by circumfixes or are they
> formed by simple suffixes from perfective verbs, therefore you must
> change the imperfective verb into its perfective counterpart? Are
> the future of the motion verbs -- like po^.jd.em 'I will go' < ist'
> (< *id.t') 'to go' -- formed by circumfixes or are they simple
> present forms of the perfecticve counterparts (since analytic
> future of imperfective forms isn't possible in these verb group)?
> Are Slovak negated verbs -- like ne.robi'.m 'I don't work' <
> robi.t' 'to work' -- formed by circumfix or are they just a
> conjunction of proclitic negating particle ne- and the respective
> personal suffix?
>
>   Therefore, I think, a circumfix may express by definition only
> one grammatical cathegory even in flexional languages. Otherwise
> this term would be excessively general, superficial, and useless.
It is a matter of the overall economy of the language.  When two
morphemes must co-occur, it will usually make sense to regard them
as a single morpheme.  Correlatives are an exception (I think).
Circumfixes seem to arise by the co-occurrence of morphemes becoming
obligatory.  The Slovak active past participles sound very like the
Germanic passive past participles in *ga-.  In Gothic, they don't
seem to be circumfixes.  In the other Germanic languages, they are
circumfixes.

I don't see any reason to declare a circumfix for the Slovak negated
verbs.

In English, 'and...respectively' probably is on the verge of being a
circumfix.  One argument against it being one is the similar usage
of 'respective' in a distributive sense.  There isn't any gain in
analysing 'not...anything' as a circumfix.

Incidentally, I think you will agree that Modern Greek has a
cicumfix.  The past indicative, both imperfective and perfective, is
now formed by the circumfix (é-)...-a-.  The augment is either
stressed or absent, in accordance with the rules of Classical Greek
accentuation.  (Unaccented word-initial vowels were dropped in the
development of Modern Greek.)

Richard.

Reply